Issues with OPLSAA force field

I calculated OPLSAA from Ligpargen python version for different small molecules for molecular crystal like PABA, ROY. The energy for PABA (para amino benzoic acid) is negative while the energy for ROY(5-methyl-2-((2-nitrophenyl)amino)thiophene-3-carbonitrile) is positive in energy minimisation step. I couldn’t figured out the issue in the force field, do I have to change manually any parameter or is there any other way to calculate OPLS. Thanks for any suggestions.

I’d usually recommend the Automated Topology Builder (ATB) over LigParGen, anecdotally.

The “energy” being positive or negative doesn’t mean much without comparison to a reference. You would have to check physical quantities, such as density vs pressure and temperature, to confirm your force field’s suitability or lack thereof.


Hi @pradipchm,

OPLS was designed with liquid structures and proteins in mind so you might get surprises when trying to simulate crystal structure of complex molecules. But I still would expect an “acceptable” description of such structure with some error compared to experimental references.

This description is totally unhelpful if you want meaningful comments on your procedure. How did you build your crystal structure? What type of minimization did you perform? Did you relax the box during minimization? Also, and quite important with OPLS, did lock hydrogen bonding distances with SHAKE? Also I read that ROY has like 13 different stable crystal structures. So again, I wonder how you got yours in the first place. PABA is way simpler as it only has a single aromatic cycle. With its two cycles, ROY seems to me way harder to work with for a start. Given the forcefield analytical form, I would guess you have overlapping atoms or highly unrelaxed molecular configuration in your structure.

I use LPG to easily get LAMMPS molecules files but as @srtee said, ATB GROMOS inputs might be a good comparison if you find ROY in the database.